X-Git-Url: http://git.mutantstargoat.com/user/nuclear/?a=blobdiff_plain;f=src%2Fscr%2Fraytrace.c;h=1593d9f7d856386325b47202981fc13ea45fa11c;hb=e5413255949140efda5ab0b1e27ec2363fcee542;hp=bc9d65431493d01302f16fbb16501f4c0c950a1c;hpb=e23b27a06797d931992ab9b8d21055f789371e2a;p=dosdemo

diff --git a/src/scr/raytrace.c b/src/scr/raytrace.c
index bc9d654..1593d9f 100644
--- a/src/scr/raytrace.c
+++ b/src/scr/raytrace.c
@@ -21,7 +21,16 @@ static struct screen scr = {
 	draw
 };
 
-static cgm_vec3 raydir[120][160];
+struct tile {
+	int x, y;
+	uint16_t *fbptr;
+};
+
+#define TILESZ		16
+#define NUM_TILES	((320 / TILESZ) * (240 / TILESZ))
+
+static cgm_vec3 raydir[240][320];
+static struct tile tiles[NUM_TILES];
 static struct rtscene scn;
 
 struct screen *raytrace_screen(void)
@@ -31,17 +40,25 @@ struct screen *raytrace_screen(void)
 
 static int init(void)
 {
-	int i, j;
+	int i, j, k;
 	float z = 1.0f / tan(cgm_deg_to_rad(25.0f));
+	struct tile *tptr = tiles;
 
-	for(i=0; i<120; i++) {
+	for(i=0; i<240; i++) {
 		cgm_vec3 *vptr = raydir[i];
-		float y = 1.0f - (float)i / 60.0f;
-		for(j=0; j<160; j++) {
-			vptr->x = ((float)j / 80.0f - 1.0f) * 1.333333f;
+		float y = 1.0f - (float)i / 120.0f;
+		for(j=0; j<320; j++) {
+			vptr->x = ((float)j / 160.0f - 1.0f) * 1.333333f;
 			vptr->y = y;
 			vptr->z = z;
 			vptr++;
+
+			if(((j & (TILESZ-1)) | (i & (TILESZ-1))) == 0) {
+				tptr->x = j;
+				tptr->y = i;
+				tptr->fbptr = fb_pixels + i * 320 + j;
+				tptr++;
+			}
 		}
 	}
 
@@ -71,33 +88,151 @@ static void start(long start_time)
 {
 }
 
-static void draw(void)
+static uint16_t INLINE rend_pixel(int x, int y)
 {
-	int i, j, r, g, b;
-	uint16_t pix, *fbptr = fb_pixels;
-
-	for(i=0; i<120; i++) {
-		for(j=0; j<160; j++) {
-			cgm_ray ray;
-			cgm_vec3 col;
-			ray.dir = raydir[i][j];
-			cgm_vcons(&ray.origin, 0, 0, -5);
-
-			if(ray_trace(&ray, &scn, 0, &col)) {
-				r = cround64(col.x * 255.0f);
-				g = cround64(col.y * 255.0f);
-				b = cround64(col.z * 255.0f);
-				if(r > 255) r = 255;
-				if(g > 255) g = 255;
-				if(b > 255) b = 255;
-				pix = PACK_RGB16(r, g, b);
-			} else {
-				pix = 0;
-			}
-			fbptr[0] = fbptr[1] = fbptr[320] = fbptr[321] = pix;
-			fbptr += 2;
+	int r, g, b;
+	cgm_ray ray;
+	cgm_vec3 col;
+
+	ray.dir = raydir[y][x];
+	cgm_vcons(&ray.origin, 0, 0, -5);
+
+	if(ray_trace(&ray, &scn, 0, &col)) {
+		r = cround64(col.x * 255.0f);
+		g = cround64(col.y * 255.0f);
+		b = cround64(col.z * 255.0f);
+		if(r > 255) r = 255;
+		if(g > 255) g = 255;
+		if(b > 255) b = 255;
+		return PACK_RGB16(r, g, b);
+	}
+	return 0;
+}
+
+#define CMPMASK		0xe79c
+static void rend_tile(uint16_t *fbptr, int x0, int y0, int tsz, int valid)
+{
+	uint16_t *cptr[4];
+	uint16_t cpix[4], tmp;
+	uint32_t pp0, pp1, pp2, pp3, *fb32;
+	int i, x1, y1, offs;
+
+	fb32 = (uint32_t*)fbptr;
+
+	if(tsz <= 2) {
+		switch(valid) {
+		case 0:
+			fbptr[1] = fbptr[320] = fbptr[321] = *fbptr;
+			break;
+		case 1:
+			fbptr[0] = fbptr[320] = fbptr[321] = fbptr[1];
+			break;
+		case 2:
+			fbptr[0] = fbptr[1] = fbptr[321] = fbptr[320];
+			break;
+		case 3:
+			fbptr[0] = fbptr[1] = fbptr[320] = fbptr[321];
+			break;
+		default:
+			printf("valid = %d\n", valid);
+			fbptr[0] = fbptr[1] = fbptr[320] = fbptr[321] = 0xff00;
+		}
+		return;
+	}
+
+	offs = tsz - 1;
+	x1 = x0 + offs;
+	y1 = y0 + offs;
+
+	cptr[0] = fbptr;
+	cptr[1] = fbptr + tsz - 1;
+	cptr[2] = fbptr + (offs << 8) + (offs << 6);
+	cptr[3] = cptr[2] + tsz - 1;
+
+	cpix[0] = valid == 0 ? *cptr[0] : rend_pixel(x0, y0);
+	cpix[1] = valid == 1 ? *cptr[1] : rend_pixel(x1, y0);
+	cpix[2] = valid == 2 ? *cptr[2] : rend_pixel(x0, y1);
+	cpix[3] = valid == 3 ? *cptr[3] : rend_pixel(x1, y1);
+
+	tmp = cpix[0] & CMPMASK;
+	if((cpix[1] & CMPMASK) != tmp) goto subdiv;
+	if((cpix[2] & CMPMASK) != tmp) goto subdiv;
+	if((cpix[3] & CMPMASK) != tmp) goto subdiv;
+
+	pp0 = cpix[0] | ((uint32_t)cpix[0] << 16);
+	pp1 = cpix[1] | ((uint32_t)cpix[1] << 16);
+	pp2 = cpix[2] | ((uint32_t)cpix[2] << 16);
+	pp3 = cpix[3] | ((uint32_t)cpix[3] << 16);
+
+	switch(tsz) {
+	case 2:
+#ifdef SUBDBG
+		pp0 = 0x18ff;
+#endif
+		fb32[0] = fb32[160] = pp0;
+		break;
+	case 4:
+#ifdef SUBDBG
+		pp0 = pp1 = pp2 = pp3 = 0x03800380;
+#endif
+		fb32[0] = fb32[160] = pp0;
+		fb32[1] = fb32[161] = pp1;
+		fb32[320] = fb32[480] = pp2;
+		fb32[321] = fb32[481] = pp3;
+		break;
+	case 8:
+#ifdef SUBDBG
+		pp1 = pp0 = pp2 = pp3 = 0xe00fe00f;
+#endif
+		fb32[0] = fb32[1] = pp0; fb32[2] = fb32[3] = pp1;
+		fb32[160] = fb32[161] = pp0; fb32[162] = fb32[163] = pp1;
+		fb32[320] = fb32[321] = pp0; fb32[322] = fb32[323] = pp1;
+		fb32[480] = fb32[481] = pp0; fb32[482] = fb32[483] = pp1;
+		fb32[640] = fb32[641] = pp2; fb32[642] = fb32[643] = pp3;
+		fb32[800] = fb32[801] = pp2; fb32[802] = fb32[803] = pp3;
+		fb32[960] = fb32[961] = pp2; fb32[962] = fb32[963] = pp3;
+		fb32[1120] = fb32[1121] = pp2; fb32[1122] = fb32[1123] = pp3;
+		break;
+
+	case 16:
+#ifdef SUBDBG
+		pp0 = 0xff00ff00;
+#endif
+		for(i=0; i<4; i++) {
+			memset16(fbptr, pp0, 16); fbptr += 320;
+			memset16(fbptr, pp0, 16); fbptr += 320;
+			memset16(fbptr, pp0, 16); fbptr += 320;
+			memset16(fbptr, pp0, 16); fbptr += 320;
 		}
-		fbptr += 320;
+		break;
+	}
+	return;
+
+subdiv:
+	*cptr[0] = cpix[0];
+	*cptr[1] = cpix[1];
+	*cptr[2] = cpix[2];
+	*cptr[3] = cpix[3];
+
+	tsz >>= 1;
+	rend_tile(fbptr, x0, y0, tsz, 0);
+	rend_tile(fbptr + tsz, x0 + tsz, y0, tsz, 1);
+	fbptr += (tsz << 8) + (tsz << 6);
+	y0 += tsz;
+	rend_tile(fbptr, x0, y0, tsz, 2);
+	rend_tile(fbptr + tsz, x0 + tsz, y0, tsz, 3);
+}
+
+static void draw(void)
+{
+	int i, j, xbound, ybound;
+	uint16_t *fbptr;
+	struct tile *tile;
+
+	tile = tiles;
+	for(i=0; i<NUM_TILES; i++) {
+		rend_tile(tile->fbptr, tile->x, tile->y, TILESZ, -1);
+		tile++;
 	}
 
 	swap_buffers(0);